stat3 inhibition overcomes temozolomide resistance in ...glioblastoma by downregulating mgmt...

Therapeutic Discovery

STAT3 Inhibition Overcomes Temozolomide Resistance inGlioblastoma by Downregulating MGMT Expression

Shinji Kohsaka1, Lei Wang1, Kazuhiro Yachi1, Roshan Mahabir1, Takuhito Narita1, Tamio Itoh3,Mishie Tanino1, Taichi Kimura1, Hiroshi Nishihara2, and Shinya Tanaka1,2

AbstractGlioblastomamultiforme (GBM) is one of themost aggressive human tumorswith apoor prognosis. Current

standard treatment includes chemotherapy with the DNA-alkylating agent temozolomide concomitant with

surgical resection and/or irradiation.However, anumber of cases are resistant to temozolomide-inducedDNA

damage due to elevated expression of the DNA repair enzyme O6-methylguanine-DNA methyltransferase

(MGMT). Here, we show that upregulation of both MGMT and STAT3 was accompanied with acquisition of

temozolomide resistance in the GBM cell line U87. Inactivation of STAT3 by inhibitor or short hairpin RNA

(shRNA) downregulated MGMT expression in GBM cell lines. MGMT upregulation was not observed by the

treatment of interleukin (IL)-6which is a strong activator of STAT3. Contrarily, forced expressedMGMT could

be downregulated by STAT3 inhibitor which was partially rescued by the proteasome inhibitor, MG132,

suggesting the STAT3-mediated posttranscriptional regulation of the protein levels of MGMT. Immunohis-

tochemical analysis of 44 malignant glioma specimens showed significant positive correlation between

expression levels of MGMT and phosphorylated STAT3 (p-STAT3; P < 0.001, r ¼ 0.58). Importantly, the

levels of both MGMT and p-STAT3 were increased in the recurrence compared with the primary lesion in

paired identical tumors of 12 cases. Finally, we showed that STAT3 inhibitor or STAT3 knockdownpotentiated

temozolomide efficacy in temozolomide-resistant GBM cell lines. Therefore, STAT3 inhibitor might be one of

the candidate reagents for combination therapy with temozolomide for patients with temozolomide-resistant

GBM. Mol Cancer Ther; 11(6); 1289–99. �2012 AACR.

IntroductionGlioma is the most common primary tumors of the

central nervous system, accounting approximately for30%, and classified into 4 clinical grades as I to IV. Themost aggressive and lethal tumors is glioblastoma multi-forme (GBM; ref. 1). The median survival of patients withGBM is less than 1 year, mainly because conventionalpostsurgical chemotherapeutic agents and irradiationexhibit limited effects (2).Temozolomide is an alkylating agent applied to malig-

nant glioma including GBM (3). Temozolomide inducesDNAmethylationofguanineatO6positionandO6-methyl-guanine incorrectly pairs with thymine and triggers

the mismatch repair (MMR) system leading to double-strand break of the genome that result in the arrest of cellcycle and induction of apoptosis (4, 5). O6-methylgua-nine-DNA methyltransferase (MGMT) removes methyl-ation from O6 position of guanine and induces temozo-lomide resistance (6). Thus, the patients with silenced ofMGMT expression through methylation of the MGMTpromoter were reported to have an improved 2-yearsurvival with temozolomide treatment together withirradiation (7). Despite pseudosubstrates of MGMT suchas O6-benzylguanine were expected to suppress resis-tance by depleting MGMT (8–10), clinical trials did notshow significant restoration of temozolomide sensitivityin patientswith temozolomide-resistantGBM(11). There-fore, other therapeutic agents which suppress MGMTexpression and sensitize the efficacy of temozolomideare highly desired (12, 13).

It iswell known that receptor tyrosine kinases includingthe EGF receptor (EGFR), platelet-derived growth factorreceptor a (PDGFRa), and the VEGF receptor (VEGFR),contribute to the growth of GBM through the activation ofRas/ERK- or phosphoinositide-3 kinase (PI3K)/AKT-dependent signaling pathways (14–16). These diversesignals converge at specific transcription factors, includ-ing STAT3 (17). STAT3 was initially identified as a signalmediator for IFN stimulation with various gene targets

Authors' Affiliations: 1Laboratory of Cancer Research, Department ofPathology, 2Department of Translational Pathology, Hokkaido UniversityGraduate School of Medicine, and 3Nakamura Memorial Hospital, Sap-poro, Japan

Note: Supplementary material for this article is available at MolecularCancer Therapeutics Online (http://mct.aacrjournals.org/).

Corresponding Author: Shinya Tanaka, Laboratory of Cancer Research,Department of Pathology, Hokkaido University Graduate School of Med-icine, N15, W7, Kita-ku, Sapporo 060-8638, Japan. Phone: 81-11-706-5052; Fax: 81-11-706-5902; E-mail: [email protected]

doi: 10.1158/1535-7163.MCT-11-0801

�2012 American Association for Cancer Research.

MolecularCancer

Therapeutics

www.aacrjournals.org 1289

on February 18, 2021. © 2012 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst April 24, 2012; DOI: 10.1158/1535-7163.MCT-11-0801

http://mct.aacrjournals.org/

controlling cell growth, and subsequently, STAT3 hasbeen shown to play an important role for cellular trans-formation (18). In fact, aberrant activation of STAT3 hasbeen identified in GBM as well as in a number of otherhuman cancers (19–21).

To date, profiling analysis has proposed molecular-based classification of GBM as proneuronal, proliferative,and mesenchymal type. Primary tumor tends to be aproneuronal type, however along with the recurrence,GBM becomes mesenchymal type with invasion andangiogenesis (22, 23). Recently, STAT3, C/EBP, bHLH-B2, RUNX1, FOSL2, and ZNF238 have been shown tocollectively control more than 74% of the mesenchymalgene expression signature gene (24). Furthermore, expres-sion of STAT3 correlates with mesenchymal differentia-tion and predicts poor clinical outcome in human glioma(24). While STAT3 is considered to be a key moleculeregulating mesenchymal phenotype, MGMT is not clas-sified into any specific subtype according to the previousmicroarray data comparing each subtype (22). Further-more, methylation status of MGMT was not associatedwith subtype (25). Several studies have shown that STAT3inhibitors suppressed STAT3 activation and expression ofits downstream target genes including cyclin D1 andVEGF (26–29). However, the relation between STAT3 andtemozolomide resistance has not yet been assessed.

In this study, to elucidate the regulatory mechanism ofMGMT and identify suppressive reagent for MGMTexpression in GBM for the increase of efficacy of temozo-lomide, we generated temozolomide-resistant cells andfound the elevation of STAT3 expression in these cells.The roles for STAT3 inGBMwere examined and the effectof STAT3 inhibitor on temozolomide-resistant GBM wasevaluated.

Materials and MethodsCells

The human GBM cell lines T98G (ATCC#CRL-1690),U138 (ATCC#HTB-16), and U87 (ATCC#HTB-14) werepurchased from American Type Culture Collection.KMG4cell linewaskindlyprovidedbyDr.KazuoTabuchi(Saga University, Saga, Japan; ref. 30). LN382, LN308, andLN235 cell lines were kindly provided by Dr. Erwin G.VanMeir (Emory University School of Medicine, Atlanta,GA; refs. 31, 32). All cell lines were cultured in Dulbecco’sModified Minimal Essential Medium (DMEM; Wako)supplemented with 10% FBS. Cell line authentication wasnot carried out by the authors within the last 6 months.

ReagentsSTAT3 inhibitor III WP1066 (Santa Cruz Biotechnolo-

gy), STAT3 inhibitor V Stattic (Santa Cruz Biotechnology),STAT3 inhibitor VI NSC 74859 (Calbiochem), Janus—activated kinase (JAK) inhibitor I (Calbiochem), METinhibitor PHA 665752 (Santa Cruz Biotechnology), andSrc-family kinase inhibitor dasatinib (LC Laboratories)were used following the manufacturer’s instruction.

STAT3 phosphorylation by IL-6 stimulationGlioma cells were cultured in DMEM supplemented

with 10% FBS for 24 hours. The cells were treated with10 ng/mL of interleukin (IL)-6 (Sigma) for 24 hours.

Preparation of retrovirus and establishment of stablecell line

For retrovirus production, the pcx4 vector system wasused (33, 34). The complete sequences of pcx4pur (puro-mycin) is available from the GenBank database(AB086386). Full-length cDNAs for human MGMT andIL-6 were subcloned into pcx4pur. Retroviruses wereobtained by using 293T cells as packaging cells and wereinfected to GBM cell lines and selected with puromycin(2 mg/mL).

Preparation of shRNA stably introduced cell lines bylentivirus system

For targeting STAT3, 3 sequences were designed by theBLOCK-It RNAi Designer (Invitrogen). The preparationof lentiviral vectors expressing human STAT3 short hair-pin RNA (shRNA) was carried out using the BLOCK-ItLentiviral RNAiExpression System (catalog no.K4934-00;Invitrogen), following the manufacturer’s instruction.Briefly, 3 pLenti6.4/shSTAT3 expression vectors contain-ing the human STAT3 shRNA-expressing cassette wereconstructed. Replication-incompetent lentivirus wasproduced by cotransfection of the pLenti6.4/shSTAT3expression vector and ViraPower PackagingMix (Invitro-gen) containing an optimized mixture of 3 packagingplasmids: pLP1, pLP2, and pLP/VSVG into 293FT cellswith Lipofectamine 2000 (Invitrogen). Viral supernatantwas harvested 48 hours after transfection, filtered througha 0.45-mmcellulose acetate filter, and frozen at�80�C. Thelentivirus-only containing pLenti6.4/U6mock vectorwasused as control. STAT3 shRNA lentiviral construct andcontrol lentiviral constructs were introduced into glio-blastoma cell lines, followed by selection with blasticidinS (5 mg/mL).

Bromodeoxyuridine assaysBrdU Labeling ELISA Kit (Roche Molecular Biochem-

icals) was used to examine the effects of temozolomide onthe proliferation of glioma cell lines. Bromodeoxyuridine(BrdUrd) is a thymidine nucleotide analogue that is incor-porated during S-phase (instead of thymidine) only in theDNA of proliferating cells. Cells were plated in 96-wellplates at a density of 2,500 cells per well and were incu-bated with DMEM supplemented with 10% FBS in thepresence or absence of temozolomide for 5 days. The rateof BrdUrd incorporation was measured by ELISA tech-nique, and the percentage of cells in S-phase was quan-titated by colorimetric evaluation (l ¼ 450 nm). Theabsorbance of cells treated with 10% FBS without temo-zolomide was assumed to be 100%, and temozolomide-induced decreases in S-phase were calculated as a per-centage of the control FBS–treated cells. Each sample was

Kohsaka et al.

Mol Cancer Ther; 11(6) June 2012 Molecular Cancer Therapeutics1290




carried out in triplicate, and the entire experiment wasrepeated twice.

MTT assayMTT assay was used to determine cell viability fol-

lowing the manufacturer’s protocols (Cayman ChemicalCompany). Cells were plated in 96-well plates at adensity of 2,500 cells per well. The cells were culturedin 100 mL of the medium with indicated concentrationof temozolomide. After 5 days, 10 mL of MTT solution(5 mg/mL) was added to each well, and the plates wereincubated at 37�C for 4 hours. Then, the media wasremoved and 100 mL of crystal dissolving solution wasadded to each well to solubilize the formazan crystals.The absorbance of the plates was measured on a micro-plate reader (Bio-Rad) at a wavelength of 570 nm. Eachsample was carried out in triplicate, and the entireexperiment was repeated twice.

Clonogenic assayClonogenic survival assays were conducted by seeding

500 cells in 6-well plates and exposing them to temozo-lomide (10–1,000 mmol/L) for 48 hours, followed by fur-ther observation for 7 to 14 days. Cell density or colonieswere assessed with crystal violet staining. Colonies ofmore than 50 cells were counted. In some clonogenicsurvival assays, the cells were cotreated with STAT3inhibitor VI.

Transwell migration assayChemotactic migration was quantified by a Boyden

chamber Transwell assay (8-mm pore size; CorningCostar) using uncoated filters. A total of 5� 104 cells weretrypsinized and introduced into the upper chamber. Afterincubation for 4 hours, using a 10% serum stimulus in thelower chamber, the remaining cells were removed fromthe upper side of the membrane by wiping, and the cellsthat had passed through to the lower side of the insertwere fixed with 100% methanol and stained with 0.04%crystal violet. The number of cells that had migrated wasthenquantifiedby counting those in 5 randomfields (�20)of each membrane. Experiments were done in triplicate.

ImmunoblottingImmunoblotting was carried out by the method de-

scribedelsewhere.Cellswere lysedwithbuffer containing0.5% NP-40, 10 mmol/L Tris-HCl (pH 7.4), 150 mmol/LNaCl, 1 mmol/L EDTA, 50 mmol/L NaF, 1 mmol/LPMFS, and 1 mmol/L Na3VO4. Proteins were subjectedto SDS-PAGE, and separated proteins were transferredto a polyvinylidene difluoride filter (Immobilon-P;Millipore). Filters were probed with antibodies obtainedfrom the following sources: anti-STAT3 monoclonal anti-bodies (mAb; Cell Signaling Technology); anti-MGMT(MT3.1) mAb and anti-actin MAb (Chemicon); anti-phos-pho-STAT3 (Tyr705) polyclonal antibody (Cell SignalingTechnology). Bound antibodies were detected withperoxidase-labeled goat antibody to mouse IgG or goat

antibody to rabbit IgG, and visualized by enhancedchemiluminescence reagents (Amersham PharmaciaBiotech).

Reverse transcriptase PCR analysisTotal RNA was isolated from cells with TRI Reagent

(Sigma) and resuspended in RNA secure resuspensionsolution (Ambion). Reverse transcription was carriedout with Superscript II RT (Invitrogen). Resultingfirst-strand cDNAwas used as a template and amplifiedby PCR using KOD-Plus DNA polymerase (Toyobo).The following primer setswere used. For STAT3, 50-CACCAA GCG AGG ACT GAG CAT-30 (sense) and 50-GCCAGA CCC AGA AGG AGA AGC-30 (antisense); forMG1MT, 50-ATG GAC AAG GAT TGT GAA-30 (sense)and 50-GTT TCG GCC AGCAGG-30 (antisense); for IL-6,50-TTCAATGAGGAGACT TGCCTG-30 (sense) and 50-ACA ACA ACA ATC TGA GGT GCC-30 (antisense); forglyceraldehyde-3-phosphate dehydrogenase (GAPDH),50-CTC ATG ACC ACA GTC CAT GC-30 (sense) and50-TTA CTC CTT GGA GGC CAT GT-30 (antisense).

Immunohistochemical analysisFormalin-fixed paraffin-embedded tissues were sec-

tioned and stained using anti-MGMT (MT3.1; Chemi-con) and anti-phospho-STAT3 (Tyr705) polyclonalantibody (Cell Signaling Technology). Scoring was asfollows: 0 (negative), 1þ (<10%–25% of cells weaklypositive), 2þ (>25% of cells weakly positive or <10%–25% of cells strongly positive), and 3þ (>25% of cellsstrongly positive).

ResultsEstablishment of temozolomide-resistant U87 GBMcell line

First, the expression levels of MGMT and temozolo-mide sensitivity in 7GBMcell lineswere investigated, and3 of them T98G, U138, and LN382, were found to highlyexpress MGMT than the other 4 cell lines KMG4, LN308,LN235, and U87 (Fig. 1A). Consistent with previousreports, MGMT-expressing cell lines exhibited a growthadvantage in the presence of temozolomide (Fig. 1B,Supplementary Fig. S1). To identify the responsible mole-cules for temozolomide resistance, comparatively temo-zolomide-sensitive U87 cell line was treated with a lowdose of temozolomide in culture media for 3 weeks andestablished temozolomide-resistant cells designated asU87R. Half maximal inhibitory concentration (IC50) forthe growth inhibition of temozolomide to U87 was lessthan 40 mmol/L, but it wasmore than 150 mmol/L inU87R(Supplementary Fig. S2A). To ensure the temozolomideresistance of U87R cells, we used clonogenic survivalassay (35), and found that the IC50 value of U87 was lessthan 10 mmol/L and that of U87R was more than400 mmol/L, suggesting that temozolomide sensitivitywas altered more than 40-fold decrease in resistant cells(Supplementary Fig. S2B). According to the increase of

Downregulation of MGMT by STAT3 Inhibitor in GBM

www.aacrjournals.org Mol Cancer Ther; 11(6) June 2012 1291




resistance, protein levels ofMGMTwere elevated inU87Rcells (Fig. 1C).

As U87R seemed to exhibit morphologic change ascuboidal cells (Fig. 1D), we examined biologic phenotypesuch as migration by Transwell assay, and found thatafter 4 hours, increased number of U87R cells (1.38-foldincrease) passed the filter compared with U87 cells(Fig. 1E). We also investigated the expression of severalcytokines, which are previously reported to promote

invasion or angiogenesis in glioma (36, 37). The expres-sion of IL-2, IL-6, and IL-10 were upregulated in U87Rcells compared with U87 cells (Fig. 1F). These resultssuggested that U87R cells possess the mesenchymal sig-nature of GBM. Therefore, we further investigated theexpression levels of STAT3, C/EBP, bHLH-B2, RUNX1,FOSL2, and ZNF238, which were reported to regulatethe mesechymal signature of GBM (24), and among themthe mRNA level of STAT3 was found to be dominantly

A

B

MGMT

Tubulin

U87 U87R

CU87 U87R

D

E

STAT 3

C/EBP

bHLH-B2

RUNX 1

FOSL 2

ZNF238

GAPDH

U87 U87R

R indicates TMZ resistance

IL-2

IL-6

IL-10

GAPDH

GF

U87 U87R

0

10

20

30

40

50

60

70

80

U87 U87R

Num

ber

of

mig

rate

d c

ells

*

MGMT

Actin

T98G LN382 U138 KMG4 LN308 LN235 U87

1 0.43 0.37 0 0 0.11 0

BrdUrd assay

0

0.2

0.4

0.6

0.8

1

1.2

T98G

LN38

2

U13

8

KMG4

LN30

8

LN23

5U87

Proliferation

N

NN

N

O

H2N

CH3

O

N

C

Temozolomide

Figure 1. Establishment oftemozolomide (TMZ)-resistant U87GBM cell line. A, immunoblotanalysis of MGMT in 7glioblastoma cell lines, T98G,LN382, U138, KMG4, LN308,LN235, and U87 (top). Actin isshown as a loading control(bottom). B, evaluation oftemozolomide resistance byBrdUrd assay. Seven glioblastomacell lineswere cultured for 5 days inthe presence of 60 mmol/L oftemozolomide. BrdUrdincorporation indicatesproliferation of cells. Error barsrepresent SD of 3 independentexperiments. Right, the structure oftemozolomide. C, immunoblotanalysis ofMGMT inU87andU87Rcells (top). Tubulin is shown as aloading control (bottom). D, phase-contrast microscopy of U87 cells(left) U87R cells (right) with �400magnification. E, evaluation ofmigration ability of U87 and U87Rby Transwell migration assay. Thenumber of migrated cells throughthe Transwell membranes with 8.0-mm pores after 4-hour incubationwas counted in 5 random fields(�20) of each membrane. Errorbars representSDof 3 independentexperiments. �,P<0.01U87 versusU87R. F, reverse transcriptasePCR (RT-PCR) analysis ofcytokine. The mRNA levels of IL-2,IL-6, and IL-10 were evaluated.GAPDH is shown as a loadingcontrol (bottom). G, RT-PCRanalysis of molecules related tomesenchymal type of GBM. ThemRNA levels of STAT3, C/EBP,bHLH-B2, RUNX1, FOSL2, andZNF238 were evaluated. GAPDH isshown as a loading control(bottom).

Kohsaka et al.





upregulated in U87R cells. Small increase of RUNX1 andalmost no significant alteration of expression of otherswere observed (Fig. 1G).

STAT3 inhibitor downregulated MGMT expressionin GBM cell linesAs concomitant upregulation of STAT3 with MGMT

was observed with acquisition of temozolomide resis-tance in U87 cells, we investigated whether STAT3 reg-ulates MGMT expression in GBM. STAT3 inhibitor VI,

NSC 74859, reduced MGMT levels in a dose-dependentmanner in T98G and U87R cells (Fig. 2A). The efficacy ofSTAT3 inhibitor was validated by the reduction of phos-phorylated levels of STAT3 (Fig. 2A). Furthermore, time-dependent suppression and recovery of MGMT levels bySTAT3 inhibitor VI were also showed (Fig. 2B). We alsoexamined effect of other STAT3 inhibitors such as inhib-itor III and inhibitor V; however, these inhibitors werecytotoxic and specific inhibition of STAT3 was notobserved (data not shown).

Figure 2. STAT3 inhibitiondownregulated MGMT expression inGBM cell lines. A, immunoblotanalysis of MGMT in T98G and U87Rtreated with STAT3 inhibitor VI (top).Dose of the STAT3 inhibitor isindicated at the top. The level ofp-STAT3 and STAT3 were alsoevaluated (second and third from thetop). Actin is shown as a loadingcontrol (bottom). B, immunoblotanalysis of MGMT in T98G treatedwith 200 mmol/L of STAT3 inhibitorVI. Duration of the treatment isindicated at the top as 0 to 48 hours.Mediumchange indicated removal ofSTAT3 inhibitor. The level of p-STAT3 was also evaluated (middle).Actin is shown as a loading control(bottom). C, immunoblot analysis ofMGMT and STAT3 in T98G, U87R,and U138 expressing STAT3 shRNAasS3sh1, S3sh2, andS3sh3 (top andmiddle). Cont indicates negativecontrol. Actin is shown as a loadingcontrol (bottom). D, immunoblotanalysis ofMGMT (top) and p-STAT3(middle) in T98G treated with JAKinhibitor I. Dose of the JAK inhibitor Iis indicated at the top. Actin is shownas a loading control (bottom).

A

B

C

(µmol/L) (-) 50 100 150 200 250

MGMT

p-STAT3

STAT3

Actin

T98G

(-) 50 100 150 200

U87R

1 0.51 0.39 0.18 0.17 0.09

1 1.03 0.98 0.43 0.13 0.03

1 0.67 0.44 0.21 0.17

1 0.85 0.90 0.73 0.20

MGMT

STAT3

Actin

831UG89T U87R

3hs3S 2hs3S 1hs3S Cont3hs3S 2hs3S 1hs3S Cont Cont S3sh1 S3sh2 S3sh3

1 0.08 0.17 0.14

1 0.10 0.19 0.11

1 0.03 0.15 0.10

1 0.05 0.08 0.06

1 0.09 0.35 0.26

1 0.15 0.28 0.28

D

(µmol/L)I

MGMT

p-STAT3

Actin

(-) 1 2 5 10 20

T98G

1 0.65 0.43 0.03 0 0

1 0.76 0.78 0.73 0.34 0.05

MGMT

p-STAT3

Actin

0 3 6 12 24 48 48 60 72 (h)

Medium change

1 0.88 0.71 0.49 0.23 0.21

1 0.86 0.81 0.89 0.22 0.24

0.24 0.81 0.80

0.25 0.93 0.95

T98G

VI

VI

VI






To exclude the possible but unpredictable effect ofSTAT3 inhibitor VI rather than STAT3 inactivation on thereduction of MGMT level, we used shRNA for STAT3 byusing lentiviral vector system and found that depletion ofSTAT3 diminished the expression of MGMT in represen-tative GBM cell lines as T98G, U87R, and U138 cells (Fig.2C). For further confirmation of the involvement ofSTAT3, we examined the effect of JAK inhibitor which isupstream of STAT3, together with the inhibitors for Src-family kinase (dasatinib), MET (PHA665752), and MEK(U0126), and among themonly JAK inhibitor could inhibitphosphorylation of STAT3 and decreased the expressionof MGMT (Fig. 2D, Supplementary Fig. S3).

Posttranscriptional regulation of MGMT expressionby STAT3

To investigate whether STAT3 regulates MGMT tran-scription, we examined the mRNA levels of MGMT aftertreatment of STAT3 inhibitor in T98G cells. mRNA levels

of MGMT were unchanged by the treatment of STAT3inhibitor VI in T98G cells (Fig. 3A). As STAT3 is oneof the downstream effectors of IL-6 (38, 39), we investi-gated whether IL-6 stimulation could activate STAT3 andinduce MGMT expression in GBM cells. While IL-6 suc-cessfully phosphorylated STAT3, induction of MGMTcould not be observed in any GBM cell line (Fig. 3B).mRNA expression levels of MGMT was mildly increasedby IL-6 overexpression in KMG4 cells whereas IL-6 over-expression did not increase the level of MGMT in U138cells (Fig. 3B).

As transcriptional regulation did not seem to beinvolved in STAT3-dependent MGMT expression, invo-lvement of the protein degradation system was analyzedand partial recovery of MGMT was observed in T98Gtreated by proteasome inhibitor, MG132 (Fig. 3C). Toconfirm nontranscriptional role for STAT3-dependent ex-pression of MGMT, we investigated whether ectopicallyexpressed MGMT could be downregulated by STAT3

A B(µmol/L)

VI(-) 100 200

MGMT

GAPDH

MGMT

IL-6

GAPDH

KMG4 U138 LN308

IL-6 MGMTMGMTIL-6

ii

Retroviralexpression contcontcont

MGMT

p-STAT3

Actin

MGMTMGMT IL-6IL-6

LN308U138KMG4iii

Retroviralexpression ContContCont

C

D

U138KMG4 LN308

IL-6

MGMT

p-STAT3

Actin

i

(-) (+) (-) (+) (-) (+)

VI

- - - +

- + ++ ++

MG132

MGMT

STAT3

p-STAT3

Actin

1 0.21 0.18 0.59

VI

LN308-pCX4-MGMTLN308

MGMT

p-STAT3

Actin

(µmol/L) (-) (-) 50 100 150 200

1 0.34 0.41 0.22 0.03

1 5.7 0.28 0.40 0.22 0.03

Figure 3. Posttranscriptional regulation of MGMT by STAT3. A, RT-PCR analysis of MGMT in T98G cells treated with 100 and 200 mmol/L of STAT3inhibitor VI for 24 hours (top). GAPDH is shown as a loading control (bottom). B, i, immunoblot analysis of MGMT and p-STAT3 in KMG4, U138, andLN308 treated with 10 nmol/L of IL-6 for 24 hours (top and middle). Actin is shown as a loading control (bottom). ii, RT-PCR analysis of MGMT and IL-6 inKMG4, U138, and LN308 cells introduced MGMT or IL-6 by retroviral vector (top and middle). GAPDH is shown as a loading control (bottom). iii, immunoblotanalysis of MGMT and p-STAT3 in KMG4, U138, and LN308 cells introduced IL-6 or MGMT by retroviral vector (top and middle). Actin is shown as aloading control (bottom). C, immunoblot analysis ofMGMTbySTAT3 inhibitor and its recovery byMG132 (top).MG132was absent (�) or present at 20 mmol/Lfor 8 hours (þ). STAT3 inhibitor VI (STAT3i-IV) was absent (�) or present at 100 mmol/L (þ) or at 200 mmol/L (þþ) for 24 hours. The level of STAT3,p-STAT3 were also evaluated (second and third from the top). Actin is shown as a loading control (bottom). D, immunoblot analysis of MGMT and p-STAT3 inLN308 and LN308 cells constituting expressing MGMT by the treatment of STAT3 inhibitor VI for 24 hours (top and middle). Dose of STAT3 inhibitor VI isindicated at the top. Actin is shown as a loading control (bottom).

Kohsaka et al.





inhibitor, and observed the reduction of MGMT proteinlevels in LN308 cells stably expressing MGMT afterSTAT3 inhibitor VI treatment (Fig. 3D).

Correlation between p-STAT3 and MGMT inmalignant glioma by immunohistochemical analysisTo investigate the presence of STAT3-dependent reg-

ulation of the levels of MGMT, immunohistochemicalanalysis was conducted by using 44 surgical specimensof human malignant glioma (Supplementary Table S1).Positivities ofMGMT andp-STAT3were evaluated as 0 to3þ (Fig. 4A), and significant correlation between the scoreof p-STAT3 and MGMT was showed (Fig. 4B). Prolifer-

ation marker KI-67/MIB-1 index did not significantlycorrelate with MGMT levels (Supplementary Table S2).No specific localization of positive tumor cells wasobserved, in terms of the tumor cells in the perivascularregion, leading edge of the lesion, or vicinity of pseudo-palisading necrosis. As we have previously reported thatlymphocytes, vascular endothelial cells, and macro-phages/microglias exhibited positivity of MGMT (13),and it was also reported that p-STAT3 staining wasconfined to the nucleus and positive for lymphocytes andvascular endothelial cells (40),we carefully scoredMGMTand p-STAT3 of tumors excluding nonneoplastic braincomponents.

Figure 4. Correlation betweenp-STAT3 and MGMT inimmunohistochemical analysisin malignant glioma. A,representative photographs ofimmunohistochemical analysis forp-STAT3 and MGMT in surgicalspecimens of malignant glioma asscore 0 to 3þ (left). Each scoring wasdetermined by combination of bothintensity and proportion ofimmunohistochemical positivityschematically showed (right), andalso described in Materials andMethods. B, correlation betweenp-STAT3 and MGMT in 44 cases ofmalignant glioma specimens. Thex- and y-axes indicate score ofpositivity of p-STAT3 and MGMT,respectively. The z-axes indicatesthe number of cases. n ¼ 44,correlation coefficient r ¼ 0.58,P < 0.001. C, immunohistochemicalanalysis of 12 cases of recurrentmalignant glioma. The score ofMGMT and p-STAT3 in primary orrecurrent tumorwas shown asgraph.Photographs of p-STAT3 andMGMTin the representative case (no. #7) inwhich primary tumor with nop-STAT3 and recurrent tumor withp-STAT3 positivity together withMGMT expression are shown.

A

B

1

1

2

2

3

3

0

MGMT

p-STAT3

5

10

n = 44

r = 0.58

P < 0.001

z

x

y

C

MGMT

0%

20%

40%

60%

80%

100%

RecurrentPrimary

2

1

0

p-STAT3

0%

20%

40%

60%

80%

100%

RecurrentPrimary

3

2

1

0

p-S

TA

T3

MG

MT

Primary Recurrent

#7

+3

0 +1

+1

p-S

TA

T3

MG

MT

Intensity

Proportion

Strong

Weak

100%25%

+2

+1

+3

+2

+3+2+10






To ensure clinical significance, we analyzed the 12sets ofmalignant glioma as primary and recurrent tumorsof identical cases. These primary tumors were surgicallyresected, temozolomide treatment was carried out, andlater on, recurrent tumors appeared and they were sur-gically removed (Supplementary Table S3). In these sets,the scores of bothMGMT and p-STAT3 were increased inthe recurrent tumors (Fig. 4C).

STAT3 inhibitor potentiated temozolomide efficacyin temozolomide-resistant GBM cell lines

To evaluate the possible combined therapy of temozo-lomide with STAT3 inhibitor, we treated the temozolo-mide-resistant cell line, T98G and U87R with temozolo-mide alone or with a combination of temozolomide andSTAT3 inhibitor VI. As T98G cells possess a remarkableamount ofMGMT, the viability of T98Gcells after 5 daysoftreatment with 200 mmol/L of temozolomide was morethan 80%, indicating the resistance to temozolomide. Incase of U87R cells, no significant growth inhibition wasobserved in range of 50 to 200mmol/L of temozolomide. Inboth T98G and U87R cells, combined treatment of temo-zolomide with STAT3 inhibitor was effective comparedwith temozolomide or STAT3 inhibitor alone (Fig. 5A).Furthermore, to ensure these results, we used the clono-genic survival assayandfound that combination treatmentwith temozolomide and STAT3 inhibitor showed syner-gistic effect of these chemicals (Fig. 5B). Especially, temo-zolomide combined with 150 or 200 mmol/L of STAT3inhibitor exhibited a strong synergistic effect. These con-centrations of STAT3 inhibitor are enough to decrease theprotein levels of MGMT as shown in Fig. 2A. Temozolo-mide did not change the expression levels of MGMT, p-STAT3, andSTAT3 (datanot shown).These results suggestthat STAT3 inhibitor attenuates temozolomide resistancein T98G cells by the decrease of protein levels of MGMT.For confirmation, we investigated whether STAT3 knock-down potentiates the therapeutic efficacy of temozolo-mide, and found that U138 cells in which STAT3 wasdepleted by shRNA were proved to be susceptible totemozolomide treatment compared with wild-type U138cells exhibiting resistance to temozolomide (Fig. 5C).

DiscussionIn this study, we showed a novel role for STAT3 in

regulation of cellular protein levels of MGMT. STAT3activation was found to be necessary for elevated levelsof MGMT in GBM. Although STAT3 is known as a tran-scriptional factor, STAT3-mediated regulation of MGMTdid not depend on its transcriptional activity. Recently, ithas been reported that active STAT3 is involved in proteinstabilization through inhibition of the ubiquitylation ofthe target molecule (41), and consistent with this report,we observed that MGMT downregulation by STAT3inhibitorwas recoveredpartially by theproteasome inhib-itor MG132 treatment. However, it should be noted thatubiquitylated levels of MGMT were unchanged after

STAT3 inhibitor treatment (data not shown). Complexformation of p-STAT3 and MGMT was not detectable byimmunoprecipitation in regular cell lysis condition (datanot shown). Therefore, other than protein degradationsuch as translational regulation of MGMT might be reg-ulated by STAT3. Especially, recent advance in researchfor miRNA (miR) raised the possibility that STAT3 mayregulate miRNAs which inhibit translation of MGMT. Ithas been reported that STAT3-regulated miR-17 played acritical role in MAP–ERK kinase (MEK) inhibitor resis-tance (42), as well as that STAT3 directly increased thelevels of miR-21 and miR-181b-1 (43). Currently onlycorrelation of some specific miRNA such as miR-21 ormiR-195 and temozolomide resistance were reported (44,45). We should address whether miRNAs are involved inthe regulation of MGMT in the future.

InMGMT-overexpressedGBMcells, the increased levelsof IL-6 and enhanced phosphorylation of STAT3 wereobserved(Fig.3BandD). Itwasreportedthat IL-6decreasedthe levels of MGMT by induction of the methylation ofMGMTpromoter inglioblastoma(46).Thus,overexpressedMGMT induces IL-6 possibly to decrease theMGMT levelsas negative feedback mechanism to maintain constantMGMT levels in the cell. As a result, STAT3 is phosphor-ylated by IL-6. Further analysis should be required for themechanism of MGMT-induced IL-6 expression.

For the relation between STAT3 status and representa-tive molecular abnormalities for glioma such as amplifi-cation/mutation of EGFR, PDGFR, or PTEN, it waspreviously reported that constitutive active STAT3 fre-quently coexpressed with EGFR in high-grade gliomas(26). In this article, we used T98G andU87 cells and EFGRstatus is wild-type in U87 cells and the expression level ofSTAT3 was low. In T98G cells, EGFR is known to beamplified and STAT3 was highly expressed (47). In caseof PTEN, both U87 and T98G have been shown to bemutated (48). Downregulation of STAT3 by PDGFRinhibitor was reported in mesangium cells in kidney(49), but there is no report for the correlation of PTEN/PDGFR and phosphorylation status of STAT3.

To ensure the biologic and clinical significance of pos-sible STAT3-mediated regulation of MGMT, correlationbetween the elevated levels of MGMT and p-STAT3 wasshowed by using human malignant glioma specimens.Cellular proliferative marker KI-67 index was not corre-lated to MGMT or p-STAT3, thus enhancement of cellcycle is not essential factor for MGMT resistance.Although the double staining of MGMT and p-STAT3 isnot technically available, the distributions of positivitiesof MGMT and p-STAT3were carefully analyzed by usingserial sections of the human GBM specimens. Furtherclinical significance of our data is shown by the increaseof the levels of both MGMT and p-STAT3 observed inrecurrent tumor comparing to the primary GBM of iden-tical patients. To date, it was reported that targetingSTAT3/JAK2 sensitizes GBM to treatment of temozolo-midewithunknownmechanism (26). In this study, STAT3seems to regulate MGMT posttranscriptionally, thus it is

Kohsaka et al.





strongly suggested that STAT3 is a therapeutic target fortemozolomide resistance in GBM.For the evaluation of possible combination therapy for

temozolomide-resistant GBM with temozolomide andSTAT3 inhibitor, we examined the efficacy of severalSTAT3 inhibitors and found that STAT3 inhibitor VIpotentiated temozolomide efficacy in temozolomide-

resistant glioma cell lines in vitro. In this study, we estab-lished temozolomide-resistant cell line by using U87which is relatively sensitive for temozolomide. Consid-ering the clinical condition, the maximum plasma con-centration of the patients administrated temozolomidewith daily dose of 150 mg/m2, was 8.22 mg/mL (42mmol/L; ref. 50). The concentration of temozolomide in

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 100 200 300 400

U138 control

U138 STAT3 sh-1

A

B

***

CViability

TMZ (µmol/L)

NS

NS

0

0.2

0.4

0.6

0.8

1

1.2

0 100 150 200

TMZ (-)

TMZ 100 µmol/L

TMZ 150 µmol/L

TMZ 200 µmol/L*

*

*

* *

*

*

*

0

0.2

0.4

0.6

0.8

1

1.2

0 100 150 200

TMZ(-)

TMZ 100 µmol/L

TMZ 150 µmol/L

TMZ 200 µmol/L

0

0.2

0.4

0.6

0.8

1

1.2

0 100 150 200

TMZ (-)

TMZ 100 µmol/L

TMZ 200 µmol/L

0

0.2

0.4

0.6

0.8

1

1.2

0 100 150 200

TMZ (-)

TMZ 100 µmol/L

TMZ 200 µmol/L

ViabilityViability

Clonogenic

survival

Clonogenic

survival

STAT3

(µmol/L)

VISTAT3

(µmol/L)

VI

STAT3

(µmol/L)

VISTAT3

(µmol/L)

VI

*

†

*

†*

*

NS NS

*

† **

##†

†

NS

*

†

†

†

NS

*

†

†

*

*

*

T98G

T98G

U87R

U87R

U138

Figure 5. STAT3 inhibitor and STAT3 knockdown potentiate temozolomide (TMZ) efficacy in temozolomide-resistant glioma cell lines. A, MTT assay of GBMtreated bySTAT3 inhibitor. T98GandU87Rcellswere cultured in the presenceor absence of indicated concentration of STAT3 inhibitor and temozolomide for5 days. Error bars represent SD of 3 independent experiments. �, P < 0.01 versus temozolomide (�); #, P < 0.01 versus temozolomide 100 mmol/L; †, P < 0.01versus temozolomide 150 mmol/L. B, clonogenic assay of GBM treated by STAT3 inhibitor. Five hundred cells of T98G andU87Rwere seeded in 6-well platesand exposed to temozolomide (0–200 mmol/L) and STAT3 inhibitor VI (0–200 mmol/L) for 48 hours. After 7 days, colonies of more than 50 cells were counted.Error bars represent SD of 3 independent experiments. �, P < 0.01 versus temozolomide 100 mmol/L; †, P < 0.01 versus temozolomide (�). C, MTT assay ofGBM in which MGMT was suppressed by shRNA. U138 and U138-shSTAT3 cells were cultured in the presence or absence of indicated concentrations oftemozolomide for 5 days. Error bars represent SD of 3 independent experiments. �, P < 0.01 versus U138 control. NS, not significant.






this experiment was approximately 40 to 150 mmol/L, thealteration of the sensitivity to temozolomide of U87 andU87R might closely mimic the clinical concentrations.Comparing the several genes expression between U87and U87R, we discovered upregulation of STAT3 expres-sion in U87R.

We also observed that JAK inhibitor I downregulatedthe expression of MGMT. However, STAT3 phosphory-lation and MGMT expression did not completely syn-chronize compared with STAT3 inhibitor VI. This mightbe because JAK inhibitor I also inhibited Jak1, Jak2, Jak3and Tyk2 and blocked STAT5 (51), and these off-targeteffects possibly inhibited the decrease of MGMT.

In this study,we couldnot conclude that combination oftemozolomide and STAT3 inhibitor VI was effective forthe regression of GBM in vivo. In fact, even after STAT3inhibitor treatment, neither inhibition of p-STAT3 norMGMTwas observed inmice xenograft (data not shown).Onepossible explanationwhy the STAT3 inhibitor didnotfunction in vivo is thequickmetabolismof STAT3 inhibitorVI in vivo. In addition, considering the short half-life oftemozolomide in vivo as about 2 hours (52, 53), continuoussupply of inhibitors for both temozolomide and STAT3should be required for the appropriate evaluation of thiscombination therapy in vivo.

It has been reported that MGMT expression was sup-pressed by IFN-b through the inactivation of transcrip-tional activity of p53 and suppressed the promoter ofMGMT (54, 55). As our findings suggested the possibleposttranscriptional suppression of MGMT protein levelsby STAT3 inhibitor, one might hypothesized that combi-nation of both IFN-b and STAT3 inhibitor might further

potentiate temozolomide efficacy hopefully applied toGBM cases in the future.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: S. Kohsaka, M. Tanino, S. TanakaDevelopment of methodology: S. Kohsaka, L. Wang, S. TanakaAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): S. Kohsaka, L. Wang, T. Narita, T. Itoh, M.Tanino, S. TanakaAnalysis and interpretation of data (e.g., statistical analysis, biostatis-tics, computational analysis): S. Kohsaka, L. Wang, M. Tanino, S. TanakaWriting, review, and/or revision of the manuscript: S. Kohsaka, R.Mahabir, S. TanakaAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): S. Kohsaka, K. Yachi, T. Narita, M.Tanino, H. Nishihara, S. TanakaStudy supervision: S. Kohsaka, T. Kimura, H. Nishihara, S. Tanaka

AcknowledgmentsThe authors thank Drs. Kazuo Tabuchi (Saga University, Saga, Japan)

and Erwin G. Van Meir (Emory University School of Medicine, Atlanta,GA) for providing cell lines; Dr. Tsuyoshi Akagi (KANResearch Institute,Inc., Kobe, Japan) for providing the plasmid; and EikoAoyanagi andMihoNodagashira for technical assistance.

Grant SupportThis work was supported in-part by Grants-in-Aid for Scientific

Research from the Ministry of Education, Culture, Sports, Science, andTechnology (MEXT) of Japan.

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received October 6, 2011; revised February 28, 2012; acceptedMarch 29,2012; published OnlineFirst April 24, 2012.

References1. Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, Burger PC, Jouvet A,

et al. The 2007 WHO classification of tumours of the central nervoussystem. Acta Neuropathol 2007;114:97–109.

2. VanMeir EG,HadjipanayisCG,NordenAD,ShuHK,WenPY,Olson JJ.Exciting new advances in neuro-oncology: the avenue to a cure formalignant glioma. CA Cancer J Clin 2010;60:166–93.

3. Mrugala MM, Chamberlain MC. Mechanisms of disease: temozolo-mide and glioblastoma–look to the future. Nat Clin Pract Oncol2008;5:476–86.

4. ChengCL, JohnsonSP,Keir ST,Quinn JA,Ali-OsmanF,SzaboC, et al.Poly(ADP-ribose) polymerase-1 inhibition reverses temozolomideresistance in a DNAmismatch repair-deficient malignant glioma xeno-graft. Mol Cancer Ther 2005;4:1364–8.

5. TrivediRN,AlmeidaKH,Fornsaglio JL,SchamusS,SobolRW. The roleof base excision repair in the sensitivity and resistance to temozolo-mide-mediated cell death. Cancer Res 2005;65:6394–400.

6. Kokkinakis DM, Bocangel DB, Schold SC, Moschel RC, Pegg AE.Thresholds of O6-alkylguanine-DNA alkyltransferase which confersignificant resistance of human glial tumor xenografts to treatmentwith 1,3-bis(2-chloroethyl)-1-nitrosourea or temozolomide. Clin Can-cer Res 2001;7:421–8.

7. Hegi ME, Diserens AC, Gorlia T, Hamou MF, de Tribolet N, Weller M,et al. MGMT gene silencing and benefit from temozolomide in glio-blastoma. N Engl J Med 2005;352:997–1003.

8. RansonM,MiddletonMR,Bridgewater J, LeeSM,DawsonM, JowleD,et al. Lomeguatrib, a potent inhibitor of O6-alkylguanine-DNA-alkyl-transferase: phase I safety, pharmacodynamic, and pharmacokinetic

trial and evaluation in combination with temozolomide in patients withadvanced solid tumors. Clin Cancer Res 2006;12:1577–84.

9. Khan O, Middleton MR. The therapeutic potential of O6-alkylguanineDNA alkyltransferase inhibitors. Expert Opin Investig Drugs 2007;16:1573–84.

10. Quinn JA, Desjardins A, Weingart J, Brem H, Dolan ME, Delaney SM,et al. Phase I trial of temozolomide plus O6-benzylguanine for patientswith recurrent or progressive malignant glioma. J Clin Oncol2005;23:7178–87.

11. Quinn JA, Jiang SX, Reardon DA, Desjardins A, Vredenburgh JJ, RichJN, et al. Phase II trial of temozolomideplus o6-benzylguanine in adultswith recurrent, temozolomide-resistant malignant glioma. J Clin Oncol2009;27:1262–7.

12. Sasai K, Akagi T, Aoyanagi E, Tabu K, Kaneko S, Tanaka S. O6-methylguanine-DNA methyltransferase is downregulated in trans-formed astrocyte cells: implications for anti-glioma therapies. MolCancer 2007;6:36.

13. Sasai K, Nodagashira M, Nishihara H, Aoyanagi E, Wang L, Katoh M,et al. Careful exclusion of non-neoplastic brain components is requiredfor an appropriate evaluation of O6-methylguanine-DNA methyltrans-ferase status in glioma: relationship between immunohistochemistryand methylation analysis. Am J Surg Pathol 2008;32:1220–7.

14. Rao RD, James CD. Altered molecular pathways in gliomas: anoverview of clinically relevant issues. Semin Oncol 2004;31:595–604.

15. Alcantara Llaguno SR, Chen J, Parada LF. Signaling in malignantastrocytomas: role of neural stemcells and its therapeutic implications.Clin Cancer Res 2009;15:7124–9.

Kohsaka et al.





16. Furnari FB, Fenton T, Bachoo RM, Mukasa A, Stommel JM, Stegh A,et al. Malignant astrocytic glioma: genetics, biology, and paths totreatment. Genes Dev 2007;21:2683–710.

17. Brantley EC, Benveniste EN. Signal transducer and activator of tran-scription-3: a molecular hub for signaling pathways in gliomas. MolCancer Res 2008;6:675–84.

18. Yu CL, Meyer DJ, Campbell GS, Larner AC, Carter-Su C, Schwartz J,et al. Enhanced DNA-binding activity of a Stat3-related protein in cellstransformed by the Src oncoprotein. Science 1995;269:81–3.

19. Bromberg J. Stat proteins and oncogenesis. J Clin Invest 2002;109:1139–42.

20. Yu H, Jove R. The STATs of cancer–new molecular targets come ofage. Nat Rev Cancer 2004;4:97–105.

21. Dasgupta A, Raychaudhuri B, Haqqi T, Prayson R, Van Meir EG,Vogelbaum M, et al. Stat3 activation is required for the growth ofU87 cell-derived tumours in mice. Eur J Cancer 2009;45:677–84.

22. Phillips HS, Kharbanda S, Chen R, Forrest WF, Soriano RH, Wu TD,et al. Molecular subclasses of high-grade glioma predict prognosis,delineate a pattern of disease progression, and resemble stages inneurogenesis. Cancer Cell 2006;9:157–73.

23. Tso CL, Shintaku P, Chen J, Liu Q, Liu J, Chen Z, et al. Primaryglioblastomas express mesenchymal stem-like properties. Mol Can-cer Res 2006;4:607–19.

24. CarroMS, LimWK, AlvarezMJ, Bollo RJ, Zhao X, Snyder EY, et al. Thetranscriptional network for mesenchymal transformation of braintumours. Nature 2010;463:318–25.

25. VerhaakRG,HoadleyKA,PurdomE,WangV,Qi Y,WilkersonMD, et al.Integrated genomic analysis identifies clinically relevant subtypes ofglioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR,and NF1. Cancer Cell 2010;17:98–110.

26. Lo HW, Cao X, Zhu H, Ali-Osman F. Constitutively activated STAT3frequently coexpresses with epidermal growth factor receptor in high-grade gliomas and targeting STAT3 sensitizes them to Iressa andalkylators. Clin Cancer Res 2008;14:6042–54.

27. Wei D, Le X, Zheng L, Wang L, Frey JA, Gao AC, et al. Stat3 activationregulates the expression of vascular endothelial growth factor andhuman pancreatic cancer angiogenesis and metastasis. Oncogene2003;22:319–29.

28. Niu G, Wright KL, Huang M, Song L, Haura E, Turkson J, et al.Constitutive Stat3 activity up-regulates VEGF expression and tumorangiogenesis. Oncogene 2002;21:2000–8.

29. Sinibaldi D, Wharton W, Turkson J, Bowman T, Pledger WJ, Jove R.Induction of p21WAF1/CIP1 and cyclin D1 expression by the Srconcoprotein in mouse fibroblasts: role of activated STAT3 signaling.Oncogene 2000;19:5419–27.

30. Wang L, Tabu K, Kimura T, Tsuda M, Linghu H, Tanino M, et al.Signaling adaptor protein Crk is indispensable for malignant featureof glioblastoma cell line KMG4. Biochem Biophys Res Commun2007;362:976–81.

31. Ishii N, Maier D, Merlo A, Tada M, Sawamura Y, Diserens AC, et al. Fre-quent co-alterations of TP53, p16/CDKN2A, p14ARF, PTEN tumor sup-pressor genes in human glioma cell lines. Brain Pathol 1999;9:469–79.

32. Albertoni M, Daub DM, Arden KC, Viars CS, Powell C, Van Meir EG.Genetic instability leads to loss of both p53 alleles in a humanglioblastoma. Oncogene 1998;16:321–6.

33. Kohsaka S, Sasai K, Takahashi K, Akagi T, TaninoM, Kimura T, et al. Apopulation of BJ fibroblasts escaped from Ras-induced senescencesusceptible to transformation. Biochem Biophys Res Commun2011;410:878–84.

34. Akagi T, Sasai K, Hanafusa H. Refractory nature of normal humandiploid fibroblastswith respect to oncogene-mediated transformation.Proc Natl Acad Sci U S A 2003;100:13567–72.

35. Hermisson M, Klumpp A, Wick W, Wischhusen J, Nagel G, Roos W,et al.O6-methylguanineDNAmethyltransferaseandp53statuspredicttemozolomide sensitivity in human malignant glioma cells. J Neuro-chem 2006;96:766–76.

36. WagnerS, StegenC,BouterfaH,HuettnerC, KerkauS,RoggendorfW,et al. Expression of matrix metalloproteinases in human glioma celllines in the presence of IL-10. J Neurooncol 1998;40:113–22.

37. Liu Q, Li G, Li R, Shen J, He Q, Deng L, et al. IL-6 promotion ofglioblastoma cell invasion and angiogenesis in U251 and T98G celllines. J Neurooncol 2010;100:165–76.

38. Grivennikov S, Karin M. Autocrine IL-6 signaling: a key event intumorigenesis? Cancer Cell 2008;13:7–9.

39. Gao SP, Mark KG, Leslie K, Pao W, Motoi N, Gerald WL, et al.Mutations in the EGFR kinase domain mediate STAT3 activation viaIL-6 production in human lung adenocarcinomas. J Clin Invest2007;117:3846–56.

40. Abou-GhazalM,YangDS,QiaoW,Reina-Ortiz C,Wei J, KongLY, et al.The incidence, correlation with tumor-infiltrating inflammation, andprognosis of phosphorylated STAT3 expression in human gliomas.Clin Cancer Res 2008;14:8228–35.

41. Jung JE, Kim HS, Lee CS, Shin YJ, Kim YN, Kang GH, et al. STAT3inhibits the degradation of HIF-1alpha by pVHL-mediated ubiquitina-tion. Exp Mol Med 2008;40:479–85.

42. Dai B, Meng J, Peyton M, Girard L, Bornmann WG, Ji L, et al. STAT3mediates resistance to MEK inhibitor through microRNA miR-17.Cancer Res 2011;71:3658–68.

43. Iliopoulos D, Jaeger SA, Hirsch HA, Bulyk ML, Struhl K. STAT3activation of miR-21 and miR-181b-1 via PTEN and CYLD are part ofthe epigenetic switch linking inflammation to cancer. Mol Cell2010;39:493–506.

44. Shi L, Chen J, Yang J, Pan T, Zhang S, Wang Z. MiR-21 protectedhuman glioblastoma U87MG cells from chemotherapeutic drug temo-zolomide induced apoptosis by decreasing Bax/Bcl-2 ratio and cas-pase-3 activity. Brain Res 2010;1352:255–64.

45. Ujifuku K, Mitsutake N, Takakura S, Matsuse M, Saenko V, Suzuki K,et al. miR-195, miR-455-3p and miR-10a(�) are implicated in acquiredtemozolomide resistance in glioblastomamultiforme cells. Cancer Lett2010;296:241–8.

46. Piperi C, Themistocleous MS, Papavassiliou GA, Farmaki E, LevidouG, Korkolopoulou P, et al. High incidence of MGMT and RARbetapromoter methylation in primary glioblastomas: association with his-topathological characteristics, inflammatory mediators and clinicaloutcome. Mol Med 2010;16:1–9.

47. Carrasco-Garcia E, SacedaM,Grasso S, Rocamora-Reverte L, CondeM, Gomez-Martinez A, et al. Small tyrosine kinase inhibitors interruptEGFR signaling by interacting with erbB3 and erbB4 in glioblastomacell lines. Exp Cell Res 2011;317:1476–89.

48. Sinn B, Tallen G, Schroeder G, Grassl B, Schulze J, Budach V, et al.Caffeine confers radiosensitization of PTEN-deficient malignant glio-ma cells by enhancing ionizing radiation-induced G1 arrest and neg-atively regulating Akt phosphorylation.MolCancer Ther 2010;9:480–8.

49. Hirai T, Masaki T, Kuratsune M, Yorioka N, Kohno N. PDGF receptortyrosine kinase inhibitor suppressesmesangial cell proliferation involv-ing STAT3 activation. Clin Exp Immunol 2006;144:353–61.

50. BealeP,Judson I,MooreS,StatkevichP,MarcoA,CutlerDL,etal. Effectof gastric pH on the relative oral bioavailability and pharmacokinetics oftemozolomide. Cancer Chemother Pharmacol 1999;44:389–94.

51. Lucet IS, Fantino E, Styles M, Bamert R, Patel O, Broughton SE, et al.The structural basis of Janus kinase 2 inhibition by a potent andspecific pan-Janus kinase inhibitor. Blood 2006;107:176–83.

52. Hammond LA, Eckardt JR, Baker SD, Eckhardt SG, DuganM, Forral K,et al. Phase I and pharmacokinetic study of temozolomide on a daily-for-5-days schedule in patients with advanced solid malignancies. JClin Oncol 1999;17:2604–13.

53. Britten CD, Rowinsky EK, Baker SD, Agarwala SS, Eckardt JR, Bar-rington R, et al. A Phase I and pharmacokinetic study of temozolomideandcisplatin in patientswith advanced solidmalignancies. ClinCancerRes 1999;5:1629–37.

54. Motomura K, Natsume A, Kishida Y, Higashi H, Kondo Y, Nakasu Y,et al. Benefits of interferon-beta and temozolomide combination ther-apy for newly diagnosed primary glioblastoma with the unmethylatedMGMT promoter: a multicenter study. Cancer 2011;117:1721–30.

55. Natsume A, Ishii D, Wakabayashi T, Tsuno T, Hatano H, Mizuno M,et al. IFN-beta down-regulates the expression of DNA repair geneMGMT and sensitizes resistant glioma cells to temozolomide. CancerRes 2005;65:7573–9.






2012;11:1289-1299. Published OnlineFirst April 24, 2012.Mol Cancer Ther Shinji Kohsaka, Lei Wang, Kazuhiro Yachi, et al. Glioblastoma by Downregulating MGMT ExpressionSTAT3 Inhibition Overcomes Temozolomide Resistance in

Updated version

10.1158/1535-7163.MCT-11-0801doi:

Access the most recent version of this article at:

Material

Supplementary

http://mct.aacrjournals.org/content/suppl/2012/04/17/1535-7163.MCT-11-0801.DC1

Access the most recent supplemental material at:

Cited articles

http://mct.aacrjournals.org/content/11/6/1289.full#ref-list-1

This article cites 54 articles, 20 of which you can access for free at:

Citing articles

http://mct.aacrjournals.org/content/11/6/1289.full#related-urls

This article has been cited by 5 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://mct.aacrjournals.org/content/11/6/1289To request permission to re-use all or part of this article, use this link



http://mct.aacrjournals.org/lookup/doi/10.1158/1535-7163.MCT-11-0801

http://mct.aacrjournals.org/content/suppl/2012/04/17/1535-7163.MCT-11-0801.DC1

http://mct.aacrjournals.org/content/11/6/1289.full#ref-list-1

http://mct.aacrjournals.org/content/11/6/1289.full#related-urls

http://mct.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://mct.aacrjournals.org/content/11/6/1289


stat3 inhibition overcomes temozolomide resistance in ...glioblastoma by downregulating mgmt...

Documents